Huang Liwei

Dr. Huang Liwei


Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
16 August 2021 in IEEE Transactions on Cybernetics
Verified
Journal Article
Closed access
MFMingsheng Fu
Mingsheng Fu
AAAnubha Agrawal
Anubha Agrawal
AIAthirai A. Irissappane
Athirai A. Irissappane
JZJie Zhang
Jie Zhang
HLHuang Liwei
Huang Liwei
HQHong Qu
Hong Qu
Deep reinforcement learning (DRL)-based recommender systems have recently come into the limelight due to their ability to optimize long-term user engagement. A significant challenge in DRL-based recommender systems is the large action space required to represent a variety of items. The large action space weakens the sampling efficiency and thereby, affects the recommendation accuracy. In this article, we propose a DRL-based method called deep hierarchical category-based recommender system (DHCRS) to handle the large action space problem. In DHCRS, categories of items are used to reconstruct the original flat action space into a two-level category-item hierarchy. DHCRS uses two deep $Q$ -networks (DQNs): 1) a high-level DQN for selecting a category and 2) a low-level DQN to choose an item in this category for the recommendation. Hence, the action space of each DQN is significantly reduced. Furthermore, the categorization of items helps capture the users’ preferences more effectively. We also propose a bidirectional category selection (BCS) technique, which explicitly considers the category-item relationships. The experiments show that DHCRS can significantly outperform state-of-the-art methods in terms of hit rate and normalized discounted cumulative gain for long-term recommendations.

23 March 2018 in IEEE Access
Verified
Journal Article
Open access
HLHuang Liwei
Huang Liwei
HQHong Qu
Hong Qu
LZLin Zuo
Lin Zuo

5 June 2014 in Birth Defects Research Part C: Embryo Today: Reviews
Verified
Journal Article
Open access
PDF file
HLHuang Liwei
Huang Liwei
JLJoshua Henry Lipschutz
Joshua Henry Lipschutz
In the past decade, cilia have been found to play important roles in renal cystogenesis. Many genes, such as PKD1 and PKD2 which, when mutated, cause autosomal dominant polycystic kidney disease (ADPKD), have been found to localize to primary cilia. The cilium functions as a sensor to transmit extracellular signals into the cell. Abnormal cilia structure and function are associated with the development of polyscystic kidney disease (PKD). Cilia assembly includes centriole migration to the apical surface of the cell, ciliary vesicle docking and fusion with the cell membrane at the intended site of cilium outgrowth, and microtubule growth from the basal body. This review summarizes the most recent advances in cilia and PKD research, with special emphasis on the mechanisms of cytoplasmic and intraciliary protein transport during ciliogenesis.Birth Defects Research (Part C) 102:174–185, 2014. ©2014 Wiley Periodicals, Inc.

26 September 2022 in IEEE Transactions on Industrial Informatics
Verified
Journal Article
Closed access
MFMingsheng Fu
Mingsheng Fu
HLHuang Liwei
Huang Liwei
ARAnanya Rao
Ananya Rao
AIAthirai A. Irissappane
Athirai A. Irissappane
JZJie Zhang
Jie Zhang
HQHong Qu
Hong Qu
Deep reinforcement learning (DRL) based recommender systems are suitable for user cold-start problems as they can capture user preferences progressively. However, most existing DRL-based recommender systems are suboptimal, since they use the same policy to suit the dynamics of different users. We reformulate recommendation as a multitask Markov Decision Process, where each task represents a set of similar users. Since similar users have closer dynamics, a task-specific policy is more effective than a single universal policy for all users. To make recommendations for cold-start users, we use a default policy to collect some initial interactions to identify the user task, after which a task-specific policy is employed. We use Q-learning to optimize our framework and consider the task uncertainty by the mutual information regarding tasks. Experiments are conducted on three real-world datasets to verify the effectiveness of our proposed framework.

1 April 2020 in Acta Biochimica et Biophysica Sinica
Verified
Journal Article
Open access
HWHua Wei
Hua Wei
HLHuang Liwei
Huang Liwei
FWFenghua Wei
Fenghua Wei
GLGuangzhi Li
Guangzhi Li
BHBin Huang
Bin Huang
CCCong Cao
Cong Cao
The occurrence and development of diabetes seriously threaten the health of patients. Therefore, the mechanism exploration of diabetes is of great significance for more effective control of this disease. In this study, we aimed to investigate the regulatory mechanism of miR-139-5p and Notch signaling pathway on liver damage and oxidative stress in diabetic mice. The mouse model of diabetes was established, and the mice were divided into normal group, model group, negative control (NC) group, miR-139-5p mimic group, miR-139-5p inhibitor group, DAPT group, and miR-139-5p inhibitor + DAPT group. The mRNA expressions of miR-139-5p, Notch1, Jagged1, and NICD1, and the protein expressions of Notch1, Jagged1, and NICD1 were detected. In addition, HepG2 cells were cultured for high glucose induction, and cell cycle distribution and apoptosis were detected by flow cytometry. The results showed that the body weights of mice in the model, NC, miR-139-5p mimic, miR-139-5p inhibitor, DAPT, and miR-139-5p inhibitor + DAPT groups were all lower than that in the normal group. Co-localization of miR-139-5p and Notch1 was observed in the fluorescence in situ hybridization assay, and miR-139-5p was found to negatively regulate Notch1. Furthermore, reduced blood glucose level and inhibited liver oxidative stress were observed in mice with miR-139-5p overexpression or DAPT treatment. DAPT treatment reversed the increase of blood glucose level and oxidative stress injury caused by miR-139-5p silencing. In conclusion, up-regulation of miR-139-5p expression can protect liver tissue from oxidative stress injury in diabetic mice, and its mechanism may be related to the inhibition of Notch signaling pathway.

2 December 2014 in Journal of Visualized Experiments
Verified
Journal Article
Open access
PDF file
HLHuang Liwei
Huang Liwei
AXAn Xiao
An Xiao
AWAndrea Wecker
Andrea Wecker
DMDaniel A. McBride
Daniel A. McBride
SCSoo Young Choi
Soo Young Choi
JLJoshua H. Lipschutz
Joshua H. Lipschutz
Polycystic kidney disease (PKD) is one of the most common causes of end-stage kidney disease, a devastating disease for which there is no cure. The molecular mechanisms leading to cyst formation in PKD remain somewhat unclear, but many genes are thought to be involved. Wnt5a is a non-canonical glycoprotein that regulates a wide range of developmental processes. Wnt5a works through the planar cell polarity (PCP) pathway that regulates oriented cell division during renal tubular cell elongation. Defects of the PCP pathway have been found to cause kidney cyst formation. Our paper describes a method for developing a zebrafish cystic kidney disease model by knockdown of the wnt5a gene with wnt5a antisense morpholino (MO) oligonucleotides. Tg(wt1b:GFP) transgenic zebrafish were used to visualize kidney structure and kidney cysts following wnt5a knockdown. Two distinct antisense MOs (AUG - and splice-site) were used and both resulted in curly tail down phenotype and cyst formation after wnt5a knockdown. Injection of mouse Wnt5a mRNA, resistant to the MOs due to a difference in primary base pair structure, rescued the abnormal phenotype, demonstrating that the phenotype was not due to “off-target” effects of the morpholino. This work supports the validity of using a zebrafish model to study wnt5a function in the kidney.